The atmospheric oxidation of diethyl ether: Chemistry of the C 2H5-O-CH(O?)CH3 radical between 218 and 335 K

Article abstract of DOI:10.1039/b706819k

The products of the Cl atom initiated oxidation of diethyl ether (DEE) were investigated at atmospheric pressure over a range of temperatures (218-335 K) and O₂ partial pressures (50-700 Torr), both in the presence and absence of NO_x. The major products observed at 298 K and below were ethyl formate and ethyl acetate, which accounted for ≈60-80% of the reacted diethyl ether. In general, the yield of ethyl formate increased with increasing temperature, with decreasing O₂ partial pressure, and upon addition of NO to the reaction mixtures. The product yield data show that thermal decomposition reaction (3), CH₃CH₂-O-CH(O ^?)CH₃ → CH₃CH₂-O-CH=O + CH₃, and reaction (6) with O₂, CH₃CH ₂-O-CH(O^?)CH₃ + O₂ → CH₃CH₂-O-C(=O)CH₃ + HO₂ are competing fates of the CH₃CH₂-O-CH(O^?) CH₃ radical, with a best estimate of k₃/k₆ ≈ 6.9 × 10²⁴ exp(-3130/T). Thermal decomposition via C-H or C-O bond cleavage are at most minor contributors to the CH₃CH ₂-O-CH(O^?)CH₃ chemistry. The data also show that the CH₃CH₂-O-CH(O^?)CH ₃ radical is subject to a chemical activation effect. When produced from the exothermic reaction of the CH₃CH₂-O-CH(OO ^?)CH₃ radical with NO, prompt decomposition via both CH₃- and probably H-elimination occur, with yields of about 40% and ≤15%, respectively. Finally, at temperatures slightly above ambient, evidence for a change in mechanism in the absence of NO_x, possibly due to chemistry involving the peroxy radical CH₃CH₂-O-CH(OO ^?)CH₃, is presented. the Owner Societies.

Full text of DOI:10.1039/b706819k

View Article Online / Journal Homepage / Table of Contents for this issue
This paper is published as part of a PCCP Themed Issue on:
Synergies between Experimental and Theoretical
Studies of Gas Phase Reactions
Guest Edited by: Paul Seakins (Leeds) and Struan Robertson (Accelrys,
Cambridge and Leeds)
Published in issue 31, 2007 of PCCP
Image reproduced by permission of Dr Taatjes and Linda A.Houston for Sandia National
Laboratories from Phys.Chem.Chem.Phys., 2007, 9, 4315-4331
Other papers published in this issue include:
Direct detection of polyynes formation from the reaction of ethynyl radical (C
CH) and allene (CH C CH
2
H) with propyne (CH –
3
C
2
2
)
Fabien Goulay, David L. Osborn, Craig A. Taatjes, Peng Zou, Giovanni Meloni and
Stephen R. Leone, Phys. Chem. Chem. Phys., 2007
DOI: 10.1039/b614502g
Master equation methods for multiple well systems: application to the 1-,2-pentyl system
Struan H. Robertson, Michael J. Pilling, Luminita C. Jitariu and Ian H. Hillier, Phys. Chem. Chem. Phys.,
2007
DOI: 10.1039/b704736c
Ab initio methods for reactive potential surfaces
Lawrence B. Harding, Stephen J. Klippenstein and Ahren W. Jasper, Phys. Chem. Chem. Phys., 2007
DOI: 10.1039/b705390h
Visit the website for both cutting edge research papers and authoritative
review articles by leaders in a range of fields
www.rsc.org/pccp

PAPER
www.rsc.org/pccp | Physical Chemistry Chemical Physics
The atmospheric oxidation of diethyl ether: chemistry of the
ꢀ
C H –O–CH(O )CH radical between 218 and 335 K
2
5
3
John J. Orlando
Received 10th May 2007, Accepted 11th June 2007
First published as an Advance Article on the web 12th July 2007
DOI: 10.1039/b706819k
The products of the Cl atom initiated oxidation of diethyl ether (DEE) were investigated at
atmospheric pressure over a range of temperatures (218–335 K) and O partial pressures (50–700
2
x
Torr), both in the presence and absence of NO . The major products observed at 298 K and
below were ethyl formate and ethyl acetate, which accounted for E60–80% of the reacted diethyl
ether. In general, the yield of ethyl formate increased with increasing temperature, with decreasing
O
2
partial pressure, and upon addition of NO to the reaction mixtures. The product yield data
ꢀ
show that thermal decomposition reaction (3), CH
3
CH
ꢀ
2
–O–CH(O )CH
3
- CH
3
CH
2
–O–CHQO
+
CH , and reaction (6) with O , CH CH –O–CH(O )CH + O - CH CH –O–C(QO)CH +
3
2
3
2
3
2
3
2
3
ꢀ
HO are competing fates of the CH CH –O–CH(O )CH radical, with a best estimate of k /k E
2
3
2
3
3
6
2
4
6
.9 ꢁ 10 exp(ꢂ3130/T). Thermal decomposition via C–H or C–O bond cleavage are at most
ꢀ
minor contributors to the CH
ꢀ
3
CH
2
–O–CH(O )CH
3
chemistry. The data also show that the
CH
3
CH
2
–O–CH(O )CH
3
radical is subject to a chemical activation effect. When produced from
ꢀ
the exothermic reaction of the CH
decomposition via both CH
3
CH
2
–O–CH(OO )CH
3
radical with NO, prompt
3
- and probably H-elimination occur, with yields of about 40% and
r15%, respectively. Finally, at temperatures slightly above ambient, evidence for a change in
mechanism in the absence of NO , possibly due to chemistry involving the peroxy radical
ꢀ
x
CH CH –O–CH(OO )CH , is presented.
2
3
3
Introduction
at the –CH
2
– groups in the parent ether:
OH þ CH3CH2OCH2CH3ðþO2Þ
Aliphatic ethers are present in the atmosphere due to their use
as fuel additives and as solvents, and thus they can contribute
to the formation of photochemical smog. Total global emis-
sions of ethers from industrial sources are believed to be about
ꢀ
!
CH3CH2OCHðOO ÞCH3
ð1Þ
ð2Þ
ꢀ
CH3CH2OCHðOO ÞCH3 þ NO
ꢂ1
ꢂ1
ꢀ
3
.5 Tg yr , with an additional 2 Tg yr arising from biomass
! CH3CH2OCHðO ÞCH3 þ NO2
1
burning. These species are generally more reactive in the
atmosphere than alkanes of similar chain length, and have
propensities for ozone production on urban and regional
scales that are similar to or larger than alkanes of similar
The observed products then originate from the decomposition
2
of 1-ethoxyethoxy or its reaction with O :
ꢀ
ꢀ
3
CH3CH2OCHðO ÞCH3 ! CH3CH2OCHð¼OÞ þ CH
2
,3
size.
ð3Þ
Diethyl ether (DEE) is among the more common of the
aliphatic ethers. Its major destruction pathway in the atmo-
sphere is via reaction with OH. With typical urban [OH] levels
ꢀ
ꢀ
CH3CH2OCHðO ÞCH3 ! CH3CH2OCð¼OÞCH3 þ H
6
ꢂ3
ꢂ11
3
ð4Þ
exceeding 10 molecule cm and kOH = 1.3 ꢁ 10
cm
ꢂ1 ꢂ1 4
molecule
s , the atmospheric residence time of diethyl
ꢀ
ꢀ
ether (and of most other aliphatic ethers) will be r1 day. The
mechanism of the atmospheric oxidation of diethyl ether has
5–7
been explored in the laboratory by three groups. The major
CH3CH2OCHðO ÞCH3 ! CH3CH2O þ CH3CHð¼OÞ
ð5Þ
product of the oxidation at 298 K in the presence of NO is
x
ꢀ
CH3CH2OCHðO ÞCH3 þ O2
ethyl formate, while lower yields of ethyl acetate and acet-
aldehyde have also been reported. The oxidation is thought to
!
CH3CH2OCð¼OÞCH3 þ HO2
ð6Þ
ꢀ
proceed predominantly via the formation of the 1-ethox-
ꢀ
Further reactions of the CH
3
radical will lead predominantly
ꢀ
yethoxy radical, CH
3
CH
2
OCH(O )CH
3
, following OH attack
to formaldehyde, while the majority of the CH
3
CH O radicals
2
generated in reaction (5) will form additional acetaldehyde.
5
Wallington and Japar observed high yields of ethyl formate
in their chamber studies of diethyl ether oxidation in air in the
Atmospheric Chemistry Division, Earth and Sun Systems Laboratory,
National Center for Atmospheric Research, PO Box 3000, Boulder,
Co 80307
presence of NO ; a 92 ꢃ 6% yield from the OH-initiated
x
oxidation and a 87 ꢃ 6% yield from the Cl atom initiated
This journal is ꢄc the Owner Societies 2007
Phys. Chem. Chem. Phys., 2007, 9, 4189–4199 | 4189

1
2,13
oxidation were reported. Ethyl acetate and acetaldehyde were
previously.
Briefly, quantification of starting material and
not detected (yields both o5%), and the authors concluded that
decomposition pathway (3) dominated the chemistry of
products is accomplished using a Fourier Transform spectro-
meter (Bomem DA3.01) operating in the infrared. The spec-
trometer is interfaced to the chamber via a set of Hanst-type
6
-ethoxyethoxy under ambient conditions. Eberhard et al. also
1
reported large yields of ethyl formate in their studies of the OH-
initiated oxidation of diethyl ether in air at 298 K, but also
detected minor yields of ethyl acetate and acetaldehyde. Their
data indicated that pathways (3), (5), and (6) were occurring
with respective yields of 66 ꢃ 14%, 8 ꢃ 2%, and 4 ꢃ 3% per
multipass optics, which provide an observational pathlength
ꢂ1
of 32.6 m. Infrared spectra covered the range 800–3900 cm
,
ꢂ1
and were obtained at a resolution of 1 cm from the co-
addition of 200 scans (acquisition time 3–4 min). The tem-
perature of the cell was regulated by flowing either ethanol
(T o 298 K) or water (T 4 298 K) from a circulating bath
through a jacket surrounding the cell. The chamber tempera-
ture was monitored using six thermocouple gauges located
along its length, and was found to be uniform along the length
of the cell to within ꢃ1 K.
1-ethoxyethoxy formed. The product yield data of Cheema
7
x
et al. in air in the presence of NO are consistent with those
reported earlier (65 ꢃ 5% ethyl formate yield, 13 ꢃ 5%
acetaldehyde yield, and 7 ꢃ 3% ethyl acetate yield from the
OH-initiated oxidation; 79 ꢃ 9% ethyl formate yield, 11 ꢃ 3%
acetaldehyde yield, 7 ꢃ 1% ethyl acetate yield from the Cl-atom
initiated oxidation), confirming the dominance of reaction (3) in
air at 298 K. However, these workers also studied the O2
dependence of the product yields, and found the data to be
inconsistent with the exclusive occurrence of reaction (3), (5),
and (6). More specifically, the yield of ethyl acetate did not
Experiments were conducted using mixtures of Cl (1.4–
2
1
5
ꢂ3
14
3.5 ꢁ 10 molecule cm ), diethyl ether (2–7 ꢁ 10 molecule
ꢂ3
14
ꢂ3
cm ), and NO (0–14 ꢁ 10 molecule cm ) in 750 ꢃ 20 Torr
of O /N diluent, at temperatures ranging from 218 to 335 K.
The O
The O
2
2
2
partial pressure was varied between 50 and 700 Torr.
was added directly to the chamber, while minor
2
change dramatically with O
2
and did not tend to zero at low O
2
components were flushed into the chamber from smaller
calibrated volumes using a flow of N . A Xe arc lamp, directed
as would be expected if reaction (6) were its sole source. Cheema
7
et al. thus postulated the occurrence of reaction (4), H-atom
2
along the length of the chamber and filtered to provide
radiation in the 240–400 nm range, was used to photolyze
2
elimination from 1-ethoxyethoxy, an O -independent source of
ethyl acetate. While not a common occurrence in unsubstituted
Cl and thus initiate the chemistry. Mixtures were photolyzed
2
alkoxy radical chemistry, H-atom elimination from alkoxy
for 4–6 periods, each of duration 20–40 s, and an infrared
spectrum was recorded after each photolysis period (with the
Xe arc lamp blocked). Quantification of diethyl ether and its
oxidation products was carried out using commercial spectral
stripping routines. The major products, ethyl formate and
ethyl acetate, were quantified under all conditions studied
using reference spectra recorded in-house at the temperature
at which the experiments were conducted. Quantification of
minor products was carried out using either reference spectra
recorded in-house at or near the temperature under study
(CH O, CH OH), via integration of spectral bands and com-
ꢀ
8–10
radicals of structure similar to 1-ethoxyethoxy, HOCH O and
2
ꢀ
CH OCH O , has also been identified and characterized.
3
2
Using structure–additivity relationships, decomposition re-
actions (3) and (4) are expected to be exothermic and roughly
7
thermoneutral, respectively, and thus likely occur with rea-
ꢂ1
7,10
sonably small barriers, r10 kcal mol or so.
Previous
work from this lab and elsewhere has shown that chemical
activation can play a role in determining product yields in
1
1
these circumstances. When produced from the exothermic
reaction of peroxy radicals with NO, alkoxy radicals are
generated with internal excitation. In cases where the barrier
to unimolecular reaction is low, this internal energy may
exceed the barrier height and allow for ‘‘prompt’’ or activated
decomposition.
2
3
parison with integrated band intensities measured near 298 K
(the peroxynitrate C H OCH(OONO )CH ), or via compar-
2
5
2
3
ison with reference spectra obtained near room temperature
(CH ONO, CH ONO , and formic acid).
Reaction of Cl atoms with diethyl ether is very rapid, k E
3
3
2
With the specific goal of investigating the effects of chemical
activation on 1-ethoxyethoxy radical chemistry and more
closely examining the origin of the ethyl acetate product and
the nature of the H-atom elimination pathway (4), we report in
this work a study of the Cl-atom initiated oxidation of diethyl
ꢂ10
3
ꢂ1 ꢂ1 14
2.5 ꢁ 10
cm molecule s , and is thus significantly more
rapid (by a factor of 3–1000) than reaction of Cl with all
identified reaction products. Thus, consumption of products
via Cl atom reaction was of minimal concern. Correction of
measured product concentrations for loss via Cl atom reaction
was not necessary (o2%) for most products (ethyl formate,
ethyl acetate, methyl nitrate, methyl nitrite, and formic acid),
while consumption of formaldehyde and methanol by Cl atom
reaction was about 7–12% at the end of a typical experiment
(approx. 50% conversion of the parent diethyl ether). Unless
otherwise stated, formaldehyde and methanol concentrations
and yields given below are those obtained after correction for Cl
atom consumption. Control experiments showed that loss of
DEE or major reaction products via either photolysis or
heterogeneous reaction on the chamber surfaces was negligible.
Chemicals used in these experiments were obtained from the
x
ether both in the presence and absence of NO , and as a
function of both the O partial pressure (50–700 Torr) and
2
temperature (218–335 K). The data clearly support the occur-
rence of activated elimination of CH – from 1-ethoxyethoxy in
3
the presence of NO , and indicate that activated H-atom
x
elimination is also likely occurring. Relative rate coefficients
for reactions (3) and (6) as a function of temperature are also
reported, along with upper limits for the rate coefficients for
reactions (4) and (5).
Experimental
Experiments were carried out in a 2 m long, 47 L evacuable
stainless steel environmental chamber that has been described
following sources: Cl (Matheson, UHP); diethyl ether, ethyl
2
formate, ethyl acetate (Sigma-Aldrich); O (US Welding); N ,
2
2
4
190 | Phys. Chem. Chem. Phys., 2007, 9, 4189–4199
This journal is ꢄc the Owner Societies 2007

(
boil-off from liquid N
materials (Cl and NO) were used as received, while liquids
diethyl ether, ethyl formate and ethyl acetate) were subjected
2
, US Welding). Gaseous starting
x
1. Experiments conducted without NO , 298 K and below
2
Products identified following the photolysis of mixtures of Cl
x
and diethyl ether in O /N diluent (i.e., in the absence of NO )
2
(
2
2
to several freeze-pump-thaw cycles before use.
at all temperatures studied (298, 269, 251, 232 and 218 K) were
ethyl formate, ethyl acetate, formaldehyde and methanol. In
addition, formic acid was also observed at 298, 269 and 251 K.
Product concentrations observed as a function of diethyl
ether consumption are shown in Fig. 1 for an experiment
carried out in 1 atm synthetic air at 269 K. The observed
products are consistent with the basics of the established
mechanism, reactions (7), (8), (11) and (12), followed by
reactions (3) through (6).
Results and discussion
In all experiments, oxidation of diethyl ether was initiated via
2
reaction with Cl atoms produced from the photolysis of Cl .
Reaction of Cl with diethyl ether can occur at two sites, with
reaction at the –CH – groups likely to dominate:
2
Cl2 ! Cl þ Cl
ð7Þ
ꢀ
2
CH3CH2OCHðOO ÞCH3
ꢀ
Cl þ CH3CH2OCH2CH3ðþO2Þ
! 2 CH CH OCHðO ÞCH þ O
ð11aÞ
3
2
3
2
ꢀ
!
CH3CH2OCHðOO ÞCH3 þ HCl
ð8Þ
ꢀ
2
CH3CH2OCHðOO ÞCH3
If it is assumed that ethyl formate and ethyl acetate are
predominantly generated following reaction (8), it follows that
!
CH3CH2OCHðOHÞCH3
5
–7
þ CH3CH2OCð¼OÞCH3 þ O2
ð11bÞ
ð12aÞ
ð12bÞ
most (480%) of the Cl atom attack occurs at the CH
However, the possibility of ethyl formate production from
reaction at the CH groups must also be considered. Cl atom
attack at the –CH groups will lead in part to the
O radical, whose likely fate is decomposi-
2
site.
ꢀ
CH3CH2OCHðOO ÞCH3 þ HO2
3
!
CH3CH2OCHðOOHÞCH3 þ O2
3
ꢀ
3 2 2 2
CH CH OCH CH
ꢀ
tion reaction (9), leading to ethyl formate, or isomerization
reaction (10):
CH3CH2OCHðOO ÞCH3 þ HO2
! CH3CH2OCð¼OÞCH3 þ H2O þ O2
ꢀ
CH3CH2OCH2CH2O ðþO2Þ
Self-reaction (11) of the 1-ethoxyethylperoxy radical (or reac-
!
CH3CH2OCH2O2 þ CH2O
ð9Þ
tion of this species with other peroxy radicals, e.g., CH O ) is
3
2
expected to lead to the formation of both molecular (ethyl
acetate, and 1-ethoxyethanol) and radical (1-ethoxyethoxy,
ꢀ
CH3CH2OCH2O2 !! CH3CH2OCHO; other products
CH
3
CH
2
OCH(O )CH
3
) products. Additional ethyl acetate
with the
ꢀ
ꢀ
CH3CH2OCH2CH2O ðþO2Þ ! CH3CHðOO ÞOCH2CH2OH
ð10Þ
can then be generated from the reaction (6) of O
2
1-ethoxyethoxy radical, and potentially from the H-atom
elimination from 1-ethoxyethoxy, reaction (4). Ethyl formate
results from methyl radical elimination from the 1-ethoxy-
ethoxy radical, reaction (3). On a molar basis, the two ester
products, ethyl formate and ethyl acetate, account for 79 ꢃ
5% of the reacted diethyl ether in the experiment shown in
1
A computational study by Ferenac et al. has shown that
5
ꢀ
isomerization of the related CH OCH CH O species is a very
3
2
2
low-barrier process, and isomerization reaction (10) of
ꢀ
CH CH OCH CH O is expected to be even more facile.
2
3
2
2
However, their computations also suggest that decomposition
ꢀ ꢀ
3 2 2 3 2 2 2
of CH OCH CH O , and hence of CH CH OCH CH O , is
also facile and cannot be ruled out as a significant process. The
lower barrier to isomerization would point to increasing
importance of isomerization (10) versus decomposition (9) as
temperature is decreased. End-products of the isomerization
are not known, but likely include multi-functional species such
as HC(O)OCH CH OH and CH CH(OH)OCH CHO (all
2
2
3
2
conditions), and CH C(O)OCH CH OH and CH CH(OOH)-
3
2
2
3
OCH CH OH (in the absence of NO only).
x
2
2
Experimental results will be presented in three sections.
Data obtained in the absence of NO (218, 232, 251, 269, and
2
98 K) will first be discussed, followed by those obtained with
NO added to the reaction mixtures (232, 251, 269, and 298 K).
Lastly, anomalous behavior observed in a limited set of
experiments carried out at 335 K will be described. In initial
discussions, it is assumed that all observed ethyl formate and
Fig. 1 Observed (uncorrected) product concentrations versus DEE
14
consumption following photolysis of a mixture of Cl
ꢂ3
2
(35 ꢁ 10
ethyl acetate derive from attack at the –CH
2
– group, reaction
14
ꢂ3
molecule cm )/DEE (7 ꢁ 10 molecule cm )/O
2
(150 Torr)/N (610
2
(
8). The effects of a possible minor contribution to ethyl
Torr) at 269 K. Filled squares–ethyl acetate; filled triangles–formalde-
hyde; filled circles–formic acid; open circles–ethyl formate; open
diamonds–methanol. Solid lines are linear least squares fits to the data.
formate production via attack at the CH3 groups will be
discussed in section 2.
This journal is ꢄc the Owner Societies 2007
Phys. Chem. Chem. Phys., 2007, 9, 4189–4199 | 4191

Fig. 1. The remaining observed products (formaldehyde,
methanol and formic acid) are all end-products of the oxida-
tion of the methyl radical product of reaction (3), e.g.,:
rather weak IR absorber and its most intense absorption
ꢂ1
feature, centered at 1745 cm , overlaps with numerous other
carbonyl-containing products generated in these experiments.
Thus, yields as high as 15% at 296 and 269 K would have been
undetectable under the experimental conditions used herein.
Conditions typically used in lower temperature experiments
CH3O2 þ CH3O2 ! CH3O þ CH3O þ O2
CH3O2 þ CH3O2 ! CH3OH þ CH2O þ O2
CH3O þ O2 ! CH2O þ HO2
(
lower diethyl ether concentrations) were such that a less
stringent upper limit, 20–25%, could be established; however,
thermodynamic considerations suggest strongly that process
(
5) will possess a higher energy barrier than reaction (3) and
thus is unlikely to occur to a significant extent, especially at
reduced temperature. Given that two acetaldehyde molecules
are generated for every occurrence of reaction (5), it seems
reasonable to conclude that this reaction contributes o10% to
the chemistry over the entire 218–298 K temperature range.
Experiments carried out in the absence of NO should be
devoid of any chemical activation effects, and thus are well
suited to the study of the chemistry of thermalized 1-ethox-
yethoxy radical, i.e., the competition between its thermal
decomposition via reactions (3), (4) and/or (5) and its reaction
HO2 þ CH2O $ HOCH2OO
HOCH2OO !! HCOOH
CH3O2 þ HO2 ! CH3OOH þ O2
Note that, for the data shown in Fig. 1, the concentrations of
the three products derived from the CH
3
radical, after correc-
tion for secondary consumption via reaction with Cl, sums to
9
2 ꢃ 10% of the ethyl formate yield. Values of 70–90% were
typical at 298, 269 and 251 K; lower values were found at
lower temperature, where ethyl formate yields were low and
co-products were often not detectable. The slight shortfall
relative to the ethyl formate yield at the higher temperatures
with O
2
, reaction (6). Thus, experiments were carried out over
partial
a range of temperatures (218–298 K), with the O
2
pressure varied between 50 and 700 Torr at each temperature.
As an example, yields of the two main products, ethyl acetate
may be the result of the formation of CH OOH and its
3
and ethyl formate, as a function of O partial pressure at 269 K
2
subsequent loss to the chamber walls.
are shown as the open symbols in Fig. 2. The entire dataset
In addition to absorption bands due to the products just
(218–298 K; 50–700 Torr O ) is summarized in a more
2
discussed, unidentified absorption features at approximately
ꢂ1
compact form, as ratios of the yield of ethyl formate to the
total (ethyl formate + ethyl acetate) yield, in Table 1a and in
Fig. 3. Note that uncertainties in individual ester yields are
typically ꢃ7% (including uncertainty in spectral analysis, and
potential systematic uncertainty in the concentration of refer-
ence samples). This leads to an uncertainty of ꢃ(0.03 to 0.05)
for the product ratios given in Table 1 (i.e., those obtained
9
50, 1050, 1130, 1230, 1345, 1380, 1740 and 1780 cm were
clearly observed in experiments conducted above 250 K. It is
possible that some of these bands belong to the expected
product of reaction (11b), 1-ethoxyethanol, CH
3 2
CH O-
CH(OH)CH . While a spectrum of this species could not be
3
found, and samples do not appear to be available commer-
cially, we note that structurally similar molecules (e.g., di-
methoxymethane, 2-ethoxyethanol) possess some bands in
common with those observed. For example, dimethoxy-
x
with and without NO ).
A more quantitative treatment of these data is given below,
but a number of initial observations can be made from the
dataset. First, the sum of the yields of the two esters is roughly
methane possesses strong absorption features at 935, 1050,
ꢂ1 16
and 1140 cm
.
On average, the identified products (ethyl
formate, ethyl acetate, formaldehyde, methanol and formic
acid) accounted for 79 ꢃ 12% of the reacted carbon at 269 K.
Mass balances at other temperature (298, 251, 232, 218 K)
were typically somewhat lower, but always in excess of 60%.
For the lower temperatures, some of the reduced mass balance
was due to the fact that some products were sometimes
(
CH OH, CH O) or always (formic acid) formed in undetect-
3 2
able amounts, particularly when ethyl formate yields were low.
Furthermore, given that it is likely that about 30–50% of the
total product in these experiments is ethyl acetate derived from
3 2
the molecular channels of reactions of CH CH O-
ꢀ
CH(OO )CH
3
radical (either directly or via secondary conver-
sion of 1-ethoxyethanol), it can be expected that the 1-
ethoxyethanol accounts for a large fraction of the remaining
carbon.
As noted in the introduction, acetaldehyde has been identi-
fied as a product of diethyl ether oxidation by some workers,
and has been attributed to the occurrence of reaction (5). No
conclusive evidence for acetaldehyde could be found in any of
the spectra obtained in this work. However, acetaldehyde is a
Fig. 2 Fractional molar yields of ethyl formate (triangles) and ethyl
acetate (circles) as a function of O
the presence (filled symbols) and absence (open) of NO
represent fits to the data using parameters given in column two
(no NO ) and column three (with NO ) of Table 2.
2
partial pressure at 269 K, both in
x
. Solid lines
x
x
4
192 | Phys. Chem. Chem. Phys., 2007, 9, 4189–4199
This journal is ꢄc the Owner Societies 2007

Table 1 (a) Ethyl formate and ethyl acetate yields from DEE oxida-
tion in the absence of NO over a range of temperatures and O partial
x
2
pressures. Data are given as Y(EF)/[Y(EF) + Y(EA)], where Y(EF)
and Y(EA) are, respectively, fractional yields of ethyl formate and
ethyl acetate. (b) Ethyl formate and ethyl acetate yields from DEE
x 2
oxidation in the presence of NO over a range of temperatures and O
partial pressures. Data are given as Y(EF)/[Y(EF) + Y(EA)], where
Y(EF) and Y(EA) are, respectively, yields of ethyl formate and ethyl
acetate
(
a)
O
2
partial pressure/
Torr
218 K 232 K 251 K 269 K 298 K 335 K
5
1
2
3
3
4
4
4
4
0
0.21
0.11
0.30
0.22
0.42
0.36
0.49
0.47
0.43
0.59
0.57
0.61
0.60
50
50
00
50
00
25
50
75
0.18
0.30
0.55
0.56
0.41
0.40
0.39
0.07
0.05
Fig. 3 Yield ratios, Y(EF)/[Y(EF) + Y(EA)], obtained in the absence
of NO as a function of O partial pressure and temperature (solid
0.25
0.23
0.53
0.51
0.55
0.55
x
2
squares, 298 K; open triangles, 269 K; filled triangles, 251 K; open
circles, 232 K; closed circles, 218 K). Solid lines, yield ratios obtained
using the parameters derived from a fit to all data obtained in the
500
600
7
0.13
0.12
0.36
0.35
00
absence of NO
obtained using parameters derived from a fit to all data obtained
between 232 and 298 K (with and without NO , column four of Table
). See text for details.
x
(column two of Table 2). Dashed lines, yield ratios
(
O
Torr
b)
x
2
partial pressure/
2
232 K 251 K 269 K 298 K 335 K
5
1
250
0
50
0.72
0.63
0.79
0.74
0.85
0.83
0.80
0.88
0.89
0.88
0.87
constant at a given temperature, but shows a slight decrease
with decreasing temperature. Secondly, at a given tempera-
3
3
4
475
500
00
50
50
0.57
0.68
0.61
2
ture, the yield of ethyl formate decreases with increasing O ,
0.78
0.76
0.85
0.83
while the yield of ethyl acetate increases, consistent with
0.87
0.86
competition between the O -dependent production of ethyl
2
acetate via reaction (6) and ethyl formate production via
0.52
0.51
6
7
00
00
0.59
0.73
0.71
thermal decomposition of 1-ethoxyethoxy radical, reaction
(
3). Third, intercepts less than unity in Fig. 3 indicate that
there is a rather larger O -independent source of ethyl acetate
in these experiments, which increases with decreasing tempera-
2
and/or 11b) as a function of temperature. Note that the
occurrence of reaction (5) will have no impact on the fits. In
initial attempts to fit the dataset of Fig. 3 to equation (A), it
was found that separation of the contribution of the three
potential components to ethyl acetate formation (the intercept
3 2
ture. This likely results from reaction of CH CH O-
ꢀ
CH(OO )CH
3
radicals with themselves or with other organic
peroxy radicals present in the system, but could also include a
contribution from reaction (4).
Assuming that the chemistry is dominated by reactions (7),
I, k
in which the ethyl acetate produced from reaction (4) is
balanced in the fits by that produced from the RO –RO
4 6 2
, and k [O ]) was not possible; essentially, a trade-off exists
(
8), (11) and (12) and reactions (3), (4) and (6), the product
ratio discussed above (and presented in Fig. 3) is determined
by the following expression:
2
2
chemistry (the parameter I in equation (A)). Thus, similar
quality fits could be achieved with k E 0 as with k approach-
4
4
YðEFÞ=½YðEFÞ þ YðEAÞꢅ ¼ ½ðS ꢂ IÞ=Sꢅ
ðAÞ
ing the value for k3.
ꢁ
½k3=ðk3 þ k4 þ k6½O2ꢅÞꢅ
However, a strong case can be made in favor of a rather
small value for k4 relative to k . Firstly, in experiments
3
where Y(EF) and Y(EA) are the fractional molar yields of
ethyl formate and ethyl acetate, respectively; S is the sum of
the fractional yields of the two ester species; I is the fractional
yield of ethyl acetate resulting from molecular channels of
peroxy radical self- and cross-reactions (the intercept of the
conducted in the presence of NO
of ethyl acetate are greatly reduced and analysis of the data
shows that k . A more stringent constraint
r (0.25 to 0.3) * k
on the magnitude of k can probably be obtained from a
consideration of the nature of the two processes, elimination
of an H-atom in reaction (4) versus elimination of a CH group
Reaction (3) will be favored for two
reasons, because it is likely more exothermic than reaction (4),
and because CH is a better ‘‘leaving group’’ than is H-atom.
x
(see Section 2 below), yields
4
3
4
ethyl acetate data shown in Fig. 2); k
3
and k
4
are the first order
is the second-
3
1
1,17,18
rate coefficients for reactions (3) and (4); and k
6
in reaction (3).
order rate coefficient for reaction (6).
Fits of expression (A) to the data can then be used to draw
conclusions regarding the relative rates of the sources of ethyl
formate (reaction (3)) and of ethyl acetate (reactions (4), (6)
3
It is thus likely that the two reactions will have similar A-
factors, but that the activation energy for reaction (4) will
This journal is ꢄc the Owner Societies 2007
Phys. Chem. Chem. Phys., 2007, 9, 4189–4199 | 4193

ꢂ3
Table 2 Rate coefficient ratio k
ture, as derived from various fits to the data
3
/k
6
(molecule cm ) versus tempera-
possible (but unlikely) effect of a non-negligible occurrence of
, a fit was carried out with the A-factor for reaction (4) set
k
4
equal to that for reaction (3), but with the activation energy
ꢂ1
Global Fit,
no NO
Global Fit,
with NO
Simultaneous
Fit of all data
a
x
b
x
c
Temp./K
increased by 0.6 kcal mol (to mimic the k r (0.25 to 0.3) *
4
2
0
20
20
k constraint discussed above). This fitting procedure yields a
3
2
2
2
2
2
98
69
51
32
18
1.53 ꢁ 10
1.40 ꢁ 10
1.89 ꢁ 10
2
3
1
1
1
1
2
9
19
19
19
18
18
24
similar expression, k /k = 9.1 ꢁ 10 exp(ꢂ2620/T) molecule
5.7 ꢁ 10
2.8 ꢁ 10
1.2 ꢁ 10
5.4 ꢁ 10
1.3 ꢁ 10
2700
3.4 ꢁ 10
1.2 ꢁ 10
3.4 ꢁ 10
6.0 ꢁ 10
2.6 ꢁ 10
9.5 ꢁ 10
4.0 ꢁ 10
6.9 ꢁ 10
3130
3
6
9
9
8
4
19
ꢂ3
cm , which returns k
3
/k
6
values that are 4–13% lower than
= 0. A more
18
the those obtained from the fit conducted with k
4
25
thorough discussion of the relative rates of processes (3), (4)
and (6) will be presented later. However, it is apparent that,
A-factor
Activation
Temp./K
a
7.0 ꢁ 10
3910
3 6
while the value of k /k changes by more than an order of
Data derived from a fit of all data (218–298 K) obtained in the
magnitude between 218 and 298 K, the activation energy for
reaction (3) is rather low for an alkoxy radical decomposition
reaction.
b
absence of NO
x
to equation (A), see text for details. Data derived
(232–298 K) to
equation (B), see text for details. Data derived from the simultaneous
from a fit of all data obtained in the presence of NO
x
c
fit of all data obtained between 232–298 K with and without NO
present, see text for details.
x
2. Experiments conducted in the presence of NO , 298 K and
x
below
Experiments were also conducted in the presence of NO
x
over
a range of temperatures (232–298 K) and O partial pressures
2
5–7
(50–700 Torr). The expected chemistry is as follows:
ꢂ1
exceed that for reaction (3), by at least 2–3 kcal mol and
ꢂ1
likely by considerably more. Note that even a 2 kcal mol
Cl þ CH3CH2OCH2CH3ðþO2Þ
difference in activation energies for the two processes implies a
ꢀ
!
CH3CH2OCHðOO ÞCH3 þ HCl
ð8Þ
ð2aÞ
ð2bÞ
ð13Þ
3
0–100 times difference in rate for the temperature range
1
investigated herein.
1,17,18
Thus, for initial consideration it was assumed that k
negligibly small. A fit of the data to equation (A) was then
conducted using the entire NO -free dataset (218 through
98 K) simultaneously, with k fixed to zero, and I and k /k
6
4
was
ꢀ
CH3CH2OCHðOO ÞCH3 þ NO
ꢀ
!
CH3CH2OCHðO ÞCH3 þ NO2
x
2
4
3
ꢀ
CH3CH2OCHðOO ÞCH3 þ NO
each parameterized by a function of the form A exp(ꢂB/T),
where A and B are fit parameters and T is the temperature.
!
CH3CH2OCHðONO ÞCH
2
3
2
4
A rate coefficient ratio k
3
/k
6
= 1.3 ꢁ 10 exp(ꢂ2690/T)
ꢂ3
molecule cm
was obtained. Calculated yield curves are
/k
are summarized in Table 2 and in Fig. 4. To examine the
ꢀ
CH3CH2OCHðOO ÞCH3 þ NO2
shown as solid lines in Fig. 3, and retrieved values for k
3
6
!
CH3CH2OCHðOONO ÞCH
2
3
ꢀ
ꢀ
CH3CH2OCHðO ÞCH3 ! CH3CH2OCHð¼OÞ þ CH3
ð3Þ
ꢀ
CH3CH2OCHðO ÞCH3 ! CH3CH2OCð¼OÞCH þ H
3
ð4Þ
ꢀ
ꢀ
CH3CH2OCHðO ÞCH3 ! CH3CH2O þ CH3CHð¼OÞ
ð5Þ
ꢀ
CH3CH2OCHðO ÞCH3 þ O2
!
CH3CH2OCð¼OÞCH þ HO2
ð6Þ
, via
3
OH radicals will also be generated in the presence of NO
reaction of HO with NO. However, box model (Acuchem)
x
1
9
2
simulations showed that only a small fraction (5–35%) of the
ether oxidation will be initiated by OH for conditions em-
ployed in this study. Products observed at all temperatures
Fig. 4 Rate coefficient ratios, k
circles – k /k values obtained using the parameters derived from a fit
to all data obtained in the absence of NO (column two of Table 2).
Open circles – k /k values obtained using the parameters derived from
a fit to all data obtained in the presence of NO (column three of Table
). Solid line – k /k values obtained using parameters derived from a
3 6
/k , plotted in Arrhenius form. Filled
3
6
x
(
highlighted in bold above) were ethyl formate, ethyl acetate,
formaldehyde, methyl nitrite, methyl nitrate, peroxynitrates
assumed to be mostly CH CH OCH(OONO )CH ], and an
3
6
x
[
2
3
6
3
2
ꢂ1
2
3
fit to all data obtained between 232 and 298 K (with and without NO
column four of Table 2).
x
,
absorption band near 1285 cm consistent with the formation
of organic nitrate(s) other than methyl nitrate.
4
194 | Phys. Chem. Chem. Phys., 2007, 9, 4189–4199
This journal is ꢄc the Owner Societies 2007

Fig. 5 Infrared spectrum attributed to the peroxynitrate species,
CH CH OCH(OONO )CH , formed following the photolysis of Cl
DEE/NO mixtures in O /N diluent at 269 K. Absorption cross
Fig. 6 Observed (uncorrected) product concentrations versus con-
14
3
2
2
3
2
/
sumption of DEE following photolysis of a mixture of Cl
ꢂ3
(32 ꢁ 10
2
ꢂ3
14
molecule cm )/DEE (6.5 ꢁ 10 molecule cm )/NO (14 ꢁ 10
14
2
2
2
sections obtained assuming quantitative conversion of DEE to the
peroxynitrate.
ꢂ3
molecule cm )/O
2 2
(150 Torr)/N (610 Torr) at 269 K. Filled
circles – ethyl formate; open squares – sum of formaldehyde, methyl
nitrite, and methyl nitrate; open diamonds, methyl nitrite; open circles
–
methyl nitrate; solid squares – formaldehyde; solid triangles – ethyl
Specific experiments were conducted to characterize and
acetate. Solid lines – linear least squares fit to the ethyl formate and
ethyl acetate data.
quantify the peroxynitrate, via the photolysis of mixtures of
2 2
Cl , diethyl ether, and NO in air:
Cl þ CH3CH2OCH2CH3ðþO2Þ
reduced temperatures were difficult to assess due to the
increased contribution of the peroxynitrate to the measured
ꢀ
!
CH3CH2OCHðOO ÞCH3 þ HCl
ð8Þ
ꢂ1
absorption near 1300 cm . These data should strictly be
ꢀ
CH3CH2OCHðOO ÞCH3 þ NO2
considered an upper limit to the actual CH
3 2
CH OCH-
!
CH3CH2OCHðOONO ÞCH
ð13Þ
(ONO )CH yield, given the possible contribution of other
3
2
3
2
species to the measured absorption. It is noteworthy that the
nitrate yield at 298 K is measurably less than that obtained in
20
A residual spectrum (obtained following subtraction of ab-
sorption features due to diethyl ether and NO ) in an experi-
2
the comparable 2-pentylperoxy + NO reaction. The low
ment at 269 K is shown in Fig. 5. Assuming stoichiometric
nitrate yield is also consistent with the findings of Eberhard
6
et al., who found no evidence for production of CH CH O-
conversion of diethyl ether to the peroxynitrate, absorption
ꢂ18
3
2
ꢂ18
2
ꢂ1
cross sections of 1.6 ꢁ 10
cm molecule and 3.4 ꢁ 10
CH(ONO )CH in their study.
2
3
2
cm molecule are obtained for the characteristic peroxyni-
ꢂ1
Observed product concentrations versus DEE consumption
ꢂ1
ꢂ1
trate features at 1295 cm
and 1717 cm , respectively.
/diethyl
(
corrected for peroxynitrate formation) for experiments
Quantification of peroxynitrate production in Cl
2
x 2
carried out at 269 K in the presence of NO at an O partial
ether/NO photolysis experiments was carried out using either
peroxynitrate standard spectra generated at specific tempera-
tures as just described, or using integrated bandstrengths
obtained using the data in Fig. 5. In all yield data reported
below, diethyl ether consumption is corrected for the forma-
tion of the peroxynitrate—that is, product yields are calcu-
lated relative to the difference between the amount of diethyl
ether consumed and the amount of peroxynitrate produced.
Following subtraction of spectra due to methyl nitrate and
the peroxynitrate, as well as other products, residual absorp-
pressure of 150 Torr (total pressure 750 Torr) are shown in
Fig. 6. Mass balance in experiments conducted at 269 K
were on average 89 ꢃ 9%, indicating that essentially all
products are identified. Mass balances at other temperatures
(
298, 251 and 232 K) were slightly lower, about 80%. Methyl
radicals formed in the presence of NOx are converted to
formaldehyde, methyl nitrite or methyl nitrate. Note that in
the data of Fig. 6, the sum of the observed concentrations of
these three species is essentially identical to that of ethyl
formate. Furthermore, under the conditions of highest NO
x
ꢂ1
tion was observed at 1285 cm . This absorption is consistent
15
ꢂ3
(
41 ꢁ 10 molecule cm ) and lowest O
of the CH ONO and CH ONO
CH O concentration by a factor of 4.5–7.5; this result, coupled
2
(50 Torr), the sum
with organic nitrate formation in the system via reaction of
peroxy radicals with NO, e.g.,
3
3
2
concentrations exceeded the
2
ꢀ
with model studies of the system, indicates that the predomi-
CH3CH2OCHðOO ÞCH3 þ NO
nant (490%) co-product of ethyl formate was methyl radical
!
CH3CH2OCHðONO ÞCH
ð2bÞ
2
3
(
from reaction (3) following initial attack at the –CH
rather than CH O itself (from reaction (9) following attack at
the –CH groups). That is, the vast majority of the ethyl
formate is derived from attack at the CH sites. The ratio of
2
groups)
ꢂ18
2
Assuming a peak absorption cross section of 1.3 ꢁ 10
cm
2
ꢂ1
ꢂ1
molecule for the 1285 cm band (on the basis of data
obtained in our laboratory), the yield of nitrate was found to
be 6 ꢃ 2% at 298 K and 8 ꢃ 3% at 269 K. Nitrate yields at
3
2
the sum of the concentrations of the three methyl-derived
This journal is ꢄc the Owner Societies 2007
Phys. Chem. Chem. Phys., 2007, 9, 4189–4199 | 4195

it is apparent from Fig. 7 that the yield of ethyl formate does
not tend to zero at low temperature and high O partial
2
pressure. This is a clear indication of a significant occurrence
of activated decomposition of 1-ethoxyethoxy to produce
ethyl formate, since the thermal decomposition should be
essentially shut off under these conditions.
A more quantitative consideration of some of these points
will now be presented. Consider a mechanism that allows for
the formation and subsequent decomposition of activated
1
-ethoxyethoxy via either CH
3
or H-atom elimination:
ꢀ
CH3CH2OCHðOO ÞCH3 þ NO
ꢀ
!
CH3CH2OCHðO ÞCH3 þ NO2
ð2Þ
ꢀ
ꢆ
ð20Þ
!
½CH3CH2OCHðO ÞCH3ꢅ þ NO2
ꢀ
CH3CH2OCHðOO ÞCH3 þ NO
Fig. 7 Yield ratios, Y(EF)/[Y(EF) + Y(EA)], obtained in the pre-
sence of NO as a function of O partial pressure and temperature
filled squares, 335 K; open circles, 298 K; filled circles, 269 K; open
!
CH3CH2OCHðONO ÞCH
ð2bÞ
2
3
x
2
(
ꢀ
ꢆ
ꢀ
½
½
CH CH2OCHðO ÞCH3ꢅ ! CH3CH2OCHð¼OÞ þ CH
squares, 251 K; filled triangles, 232 K). Solid line, yield ratios obtained
using the parameters derived from a fit to all data obtained in the
3
3
0
ð3 Þ
presence of NO
x
(column three of Table 2). Dashed line, yield ratios
obtained using parameters derived from a fit to all data obtained
ꢀ
ꢆ
CH CH2OCHðO ÞCH3ꢅ ! CH3CH2OCð¼OÞCH3 þ H
3
x
between 232 and 298 K (with and without NO , column four of
0
ð4 Þ
Table 2). See text for details.
ꢀ
ꢀ
3
CH3CH2OCHðO ÞCH3 ! CH3CH2OCHð¼OÞ þ CH
products to the ethyl formate concentration was typically
5–105% at 298 and 269 K. Ratios at lower temperature were
ð3Þ
8
generally lower and more scattered, owing largely to difficulty
in detecting methyl nitrite under conditions where the ethyl
ꢀ
CH3CH2OCHðO ÞCH3 ! CH3CH2OCð¼OÞCH3 þ H
ð4Þ
formate yield is low. The complexity of the spectra in the
ꢂ1
1
700–1800 cm region, due to the presence of the two esters,
ꢀ
CH3CH2OCHðO ÞCH3 þ O2
formaldehyde and the peroxynitrate made identification and
quantification of acetaldehyde difficult. Nevertheless, no
evidence for this species was observed in any of the experi-
!
CH3CH2OCð¼OÞCH3 þ HO2
ð6Þ
From this mechanism, the following expression can be derived
to describe the product ratio:
ments conducted with NO present, and a yield of o20% is
x
estimated.
YðEFÞ=½YðEFÞ þ YðEAÞꢅ ¼ C þ ð1 ꢂ C ꢂ DÞ
ðBÞ
Experiments with NO
x
present were carried out over a range
partial
of temperatures (232, 251, 269, and 298 K) and O
2
ꢁ
k3=ð½k3 þ k4 þ k6½O2ꢅÞ
pressures (50–700 Torr), largely to examine effects of chemical
activation on the observed product yields. Ethyl formate and
ethyl acetate yields were again used as the main diagnostics. As
an example, molar yields of these two products are shown as a
where Y(EF) and Y(EA) are the fractional molar yields of
ethyl formate and ethyl acetate, respectively; C is the fraction
of 1-ethoxyethoxy radicals decomposing via activated process
0
function of O
2
partial pressure at 269 K in Fig. 2. The entire
(3 ), D is the fraction of 1-ethoxyethoxy radicals decomposing
0
dataset is shown in condensed form (as the ratio of the ethyl
formate yield to the sum of the yields of the two esters) in Fig. 7.
As was the case in the absence of NO, the yield of ethyl
via activated process (4 ); and k , k , and k are the rate
3
4
6
coefficients for the corresponding thermal reactions (3), (4),
and (6). Again, a minor occurrence of reaction (5) will have no
influence on the results.
formate decreases with increasing O while the yield of ethyl
2
acetate increases, consistent with a competition between
In an initial analysis, k was assumed to be negligibly small,
4
1
-ethoxyethoxy decomposition reaction (3) and its reaction
6) with O to generate ethyl acetate. Also, it is seen that the
yield ratio does not reach unity at high temperature and low
partial pressure, indicating the formation of a small
10–15%) yield of ethyl acetate under these conditions. Thus,
and expression (B) was fit to all the data in Fig. 7 with C, D,
and the A-factor (A) and activation temperature (B) for the
rate coefficient ratio k /k = A exp(ꢂB/T) as fit parameters.
(
2
3
6
O
2
Best fit, shown as the solid lines in Fig. 7 and summarized in
(
Table 2 and Fig. 4, was obtained with C = 0.47, D = 0.12,
2
5
ꢂ3
there is clearly a source of ethyl acetate that does not result
and k /k = 7.0 ꢁ 10 exp (ꢂ3910/T) molecule cm
.
3
6
from the reaction (6) of 1-ethoxyethoxy with O , an observa-
2
However, it was also found that equally good fits could be
7
tion that is consistent with the observations of Cheema et al.,
obtained by increasing the value of k while simultaneously
4
and with their conclusion that H-atom elimination is a sig-
decreasing the value of D—i.e., it is not possible to distinguish
nificant fate of 1-ethoxyethoxy in the presence of NO . Lastly,
x
between thermal and activated H-atom loss from the
4
196 | Phys. Chem. Chem. Phys., 2007, 9, 4189–4199
This journal is ꢄc the Owner Societies 2007

1
-ethoxyethoxy radical. As an extreme case, a fit was con-
thermodynamic driving force to ester formation for the ether-
derived alkoxy radicals.
ducted with the value of D fixed to zero; it was determined that
good fits could be obtained with values for k approximately
4
(c) Typically, rate coefficients for reaction of alkoxy radicals
ꢂ14
3
cm molecule s , implying that the
ꢂ1 ꢂ1 11
2
5–30% those for k , independent of temperature. As dis-
3
with O are E10
2
cussed earlier, values for k are unlikely to be this high and the
4
decomposition rate coefficient for 1-ethoxyethoxy is k E 2 ꢁ
3
6 4 ꢂ1
fit obtained with k E 0 is preferred.
10 at 298 K and E5 ꢁ 10 s at 220 K. Given that reaction
4
It is clear from Fig. 4 that quantitative differences exist at
(6) is likely more exothermic (and thus possibly faster) than
low temperature between the k
without NO present. Because there is no obvious reason to
prefer either dataset, it was thought that the best values for k
could be obtained by merging the two datasets. In order not
to excessively weight the NO -free data, the 218 K data
obtained only with no NO present) were not included in
3
/k
6
ratios obtained with and
typical alkoxy reactions with O
represent lower limits.
2 3
, these values for k likely
x
3
/
(d) Activated decomposition of 1-ethoxyethoxy via process
0
k
6
(3 ) is clearly occurring in the presence of NO , and accounts
x
x
for about E30–50% of the chemistry of the radicals generated
from the reaction of the 1-ethoxyethylperoxy radical with NO.
7
(e) In keeping with the findings of Cheema et al., there is
(
x
the fit. The remainder of the dataset (data obtained at four
temperatures, with and without NO ) was then fit collectively,
x
clearly a source of ethyl acetate other than reaction (6) in the
using six fit parameters: Two fit parameters were used to
NO -based experiments—i.e., a source that does not depend
x
represent the intercept of the NO -free traces in Arrhenius
x
on O . This may arise from a small (10–15%) occurrence of
2
0
form, two parameters were used to represent the ratio k
3
/k
6
in
activated decomposition of 1-ethoxyethoxy via (4 ), or from a
Arrhenius form, and two parameters (C and D referred to
minor occurrence of reaction (4), or some combination of the
two. On the basis of thermodynamics arguments, it seems
above) were used to represent the activated decomposition of
0
0
1
-ethoxyethoxy via processes (3 ) and (4 ) in the presence of
more likely that activated decomposition provides the O
independent source of ethyl acetate.
2
-
NO
2
x
4
. Best fit under this scenario was found with k
ꢂ3
3
/k
6
= 6.9 ꢁ
1
0
exp(ꢂ3130/T) molecule cm , and with C = 0.36 and
(f) Ethyl formate will be the major product obtained from
diethyl ether oxidation over the full range of conditions
encountered in the troposphere, even in the coldest regions
D = 0.15. Comparisons of this fit (dashed lines) with the yield
data are shown in Fig. 3 and 7, while the k /k data are
3
6
1
8
summarized in Table 2 and in Fig. 4. This fit (in which the two
datasets are combined) provides a reasonable representation
of the data, with 44 of the 48 yield ratio data points fit to
within the uncertainty in their measured value. Largely via
visual inspection of fits conducted on the entire dataset with
the activation temperature systematically varied around its
of the upper troposphere (e.g., 12 km, 220 K, [O ] = 10
2
ꢂ3
molecule cm ). This product will be formed from both
thermal and activated decomposition of 1-ethoxyethoxy.
However, there will also be a minor yield of ethyl acetate
under all tropospheric conditions, as a result of activated
H-atom elimination (or perhaps a minor occurrence of
reaction (4)).
3 6
central value of 3130 K, uncertainties in the k /k activation
temperature are estimated at ꢃ600 K, with corresponding
variations in the A-factor of a factor of 12. Lastly, uncertainty
due to the possibility of a small contribution from reaction (9),
(g) The acetaldehyde yield appears to be rather small, less
than 20% over the range of conditions studied. Given that two
acetaldehydes result from each occurrence of reaction (5), it
ethyl formate production via initial attack at the CH site, was
3
can be concluded k5 o 0.1 * k . The upper limit for the
3
considered. Recall from earlier in this section that this process
contributes at most 10% to the chemistry. Extensive examina-
tion and re-fitting of the combined dataset, with ethyl formate
yields appropriately adjusted, revealed that this reaction could
introduce a high bias (10–45%, depending on the form of
acetaldehyde reported here is consistent with yields reported in
–7
previous studies at 298 K.
5
3
.
High temperature (335 K) experiments, with and without
NOx
correction applied) in retrieved k
of potential effects on the data obtained in the absence of NO
Thus, to allow for this possibility, an asymmetric uncertainty
3 6
/k values, largely the result
x
.
A limited set of experiments was carried out at 335 K in both
the presence and absence of NO
pressures. In the presence of NO
x
over a range of O
2
partial
+
600
in the activation temperature, 3130ꢂ800 K, is adopted.
Despite these uncertainties, a number of quantitative con-
clusions can be drawn regarding the chemistry of the
x
, the results obtained were
consistent with expectations based on the lower temperature
data. The total yield of the two ester products (78 ꢃ 4%) was
high and accounted for the majority of the oxidized DEE.
Individual yields (ethyl formate, 67 ꢃ 5%; ethyl acetate, 11 ꢃ
3%) of the two ester products were found to be independent of
1
-ethoxyethoxy radical under atmospheric conditions:
a) The rate coefficient ratio k /k varies with temperature,
(
3
6
2
0
ꢂ3
decreasing from a value of about 2 ꢁ 10 molecule cm at
1
8
ꢂ3
2
98 K to less than E6 ꢁ 10 molecule cm at 218 K. The
2
O , within experimental uncertainty, indicating that at ele-
data imply a difference in activation energies between reaction
vated temperature the dissociation process (3) dominates the
chemistry of thermalized 1-ethoxyethoxy radicals. At 335 K,
+
1.2
ꢂ1
.
(
3) and reaction (6) of 6.2
/
ꢂ1.6 kcal mol
(
b) Under the assumption of the existence of a small (E0–1
kcal mol ) barrier to reaction (6), typical for alkoxy radical
reaction (6) with O
only source of ethyl acetate in the system is via the
(O -independent) activated elimination of H-atom or from a
2
no longer contributes significantly and the
ꢂ1
1
1
reactions with O
2
,
the barrier to methyl radical elimination
+1.7
2
from 1-ethoxyethoxy radical is then likely 6.7 ꢂ2.1 kcal
ꢂ1
mol . This is a very small activation barrier for an alkoxy
/
minor occurrence of thermal reaction (4).
Results obtained in the absence of NOx were, however,
more intriguing. The yields of ethyl formate and ethyl acetate
radical C–C bond scission reaction, and reflects the strong
This journal is ꢄc the Owner Societies 2007
Phys. Chem. Chem. Phys., 2007, 9, 4189–4199 | 4197

were both found to be independent of O
2
partial pressure,
there is no change in relative yields of the two ester products,
there should be no significant change in the peroxy radical
distribution and hence no change in the bimolecular chemistry
associated with the peroxy species. Finally, the apparent
strong temperature dependence to the change in mechanism
is perhaps indicative of the participation of a unimolecular
reaction.
again indicating the lack of importance of reaction (6) at
elevated temperatures and the predominance of reaction (3)
in the 1-ethoxyethoxy chemistry. However, absolute yields of
the two esters (25 ꢃ 2% for ethyl formate, and 19 ꢃ 2% for
ethyl acetate) were considerably reduced compared to those
found at lower temperatures. Following subtraction of the
absorption features due to DEE and the two ester products, a
A possible peroxy radical reaction to consider is H-atom
transfer via a six-membered ring transition state:
very complex residual absorption spectrum (peaks at 940, 1014,
ꢂ1
1075, 1120, 1160, 1245, 1340, 1385, 1740, and 1780 cm
)
ꢀ
ꢀ
CH3CH2OCHðOO ÞCH3 ! CH3CH OCHðOOHÞCH3
ð14Þ
remained. While some of these absorption features probably
correspond to those observed in the absence of NO at lower
x
temperatures, there were clearly new features present as
The analogous reaction, often designated RO
2
- QOOH, is
well known in alkylperoxy radical chemistry at elevated
well, in particular those at 940, 1014, 1075, 1250 and possibly
ꢂ1
1
775 cm
.
2
3
temperatures (see, for example, Taatjes and references there-
in). Furthermore, the analogous rearrangement reaction has
been shown to be of importance in the chemistry of dimethyl
ether (DME) at high temperature and/or low pressure (via a
A single experiment was also carried out in 1 atm air at
55 K. Under these conditions, the ethyl formate yield was
3
further reduced (r20%), while the ethyl acetate yield was not
discernible due to spectral interferences. The absorption fea-
2
4–33
ꢂ1
chemically activated process),
and it seems likely that
tures at 940, 1014, 1075, and 1250 cm were even more
isomerization will be energetically more favorable for the more
substituted radicals derived from DEE or DIPE. A likely fate
prominent at this elevated temperature. The identity and
origin of the species contributing to this residual spectrum
are unknown at this time. The observation of a different
of the rearrangement product of reaction (14) is O addition,
2
leading to the formation of an array of multi-functional
products.
x
mechanism in the absence of NO compared to its presence,
2
1
however, is reminiscent of the work of Collins et al., who
recently published a study of diisopropyl ether (DIPE) oxida-
tion at 298 K. These authors reported near-quantitative for-
mation of isopropyl acetate and formaldehyde in the presence
To be of importance in this work, the thermal isomerization
process (14) must occur with a barrier of less than about 17–18
ꢂ1
kcal mol . Studies of some related reactions—e.g., the 1,4-H
3
4
2
2
ꢀ
atom shift in O
2
CH
CH
2
–CHQO, and the 1,6-H atom shift in
x
of NO , consistent with previous studies and consistent with
ꢀ
35
CH CH(OH)CH
3
2
2
CH(OO )CH
3
2 2
or HOCH CH CH2-
the expected atmospheric peroxy and alkoxy radical chemistry:
ꢀ
36
3
CH(OO )CH —have generated barrier heights of roughly
ꢂ1
ꢀ
ðCH3Þ CðOO ÞOCHðCH3Þ þ NO
2
2
the required magnitude (14.5–19 kcal mol ). On the other
ꢂ1
ꢀ
!
ðCH3ÞCðO ÞOCHðCH3Þ þ NO2
2
hand, estimates of the barrier (18–24 kcal mol ) and/or rate
ꢀ
27,28,30,32
for the 1,5-H shift in the DME-derived CH OCH OO radi-
3
2
ꢀ
ðCH3Þ CðO ÞOCHðCH3Þ ! CH3Cð¼OÞOCHðCH3Þ þ CH3
cal
have generally shown this process to be too slow to
2
2
2
be of importance near ambient temperature, and it is ques-
tionable as to whether the rearrangement process would be
sufficiently enhanced in the DEE or DIPE cases to become
important. Obviously, definitive conclusions regarding the
observed change in mechanism cannot be made at this time.
Nonetheless, this work and the study of Collins et al. suggest
the occurrence of unknown peroxy radical chemistry at or near
298 K. Further theoretical and experimental studies of these
substituted ether-derived peroxy species are clearly warranted,
particularly given that their lifetimes may be greater than 1 s
However, in the absence of NO , yields of isopropyl acetate
x
and formaldehyde were significantly reduced and acetone
2
1
became a major oxidation product. Collins et al. suggested
ꢀ
that, for thermalized (CH
3
)
2
C(O )OCH(CH
3
)
2
radicals gener-
ated from peroxy radical self- and cross-reactions, a 1,4-H
shift
ꢀ
ꢀ
ðCH3Þ CðO ÞOCHðCH3Þ ! CH3CðOHÞOC ðCH3Þ
2
2
2
was competing with the CH
3
-elimination reaction shown
even in reasonably polluted (relatively high-NO
environments.
x
) urban
above, and was accounting for at least a portion of the
observed acetone. It seems unlikely, however, that such an
isomerization, occurring via a five-membered ring and likely
occurring through a measurable barrier, would compete with
what should be a very rapid C–C bond cleavage reaction.
It is tempting to consider the possibility that the change in
mechanism observed in this work, and perhaps in the work of
Summary
The Cl atom-initiated oxidation of diethyl ether has been
studied under conditions relevant to the lower atmosphere,
Collins et al. as well, originates from (thermal) unimolecular
ꢀ
x
both in the presence and absence of NO . Ethyl formate and
reactions of the peroxy radical, CH
ꢀ
3
CH
2
OCH(OO )CH
3
from
ethyl acetate are the major products observed under most
conditions. The data show that decomposition via C–C bond
cleavage, reaction (3), is the major fate of thermalized
DEE or (CH
3
)
2
CHOC(OO )(CH
3
)
2
from DIPE. Note that the
lifetime of the DEE-derived peroxy species in our chamber
ꢀ
system increases from less than 1 ms in the presence of NO to
x
1-ethoxyethoxy radical, CH CH OCH(O )CH , throughout
3
2
3
hundreds of ms in its absence, thus potentially allowing for the
occurrence of reasonably slow processes. Furthermore, as
the troposphere. The barrier to this decomposition process is
+
1.7
/
ꢂ1
about 6.7
kcal mol . When formed from the
ꢂ2.1
4
198 | Phys. Chem. Chem. Phys., 2007, 9, 4189–4199
This journal is ꢄc the Owner Societies 2007

exothermic reaction of the corresponding peroxy radical with
8 B. Veyret, P. Roussel and R. Lesclaux, Int. J. Chem. Kinet., 1984,
6, 1599.
9
1
M. E. Jenkin, G. D. Hayman, T. J. Wallington, M. J. Hurley, J. C.
NO, decomposition of 1-ethoxyethoxy via an activated process
0
(
3 ) is clearly occurring, and accounts for about E30–50% of
the chemistry of the radicals generated in reaction (2). In
Ball, O. J. Nielsen and T. Ellermann, J. Phys. Chem., 1993, 97,
1
1712.
10 E. Henon, F. Bohr, N. Sokolowski-Gomez and F. Caralp, Phys.
Chem. Chem. Phys., 2003, 5, 5431.
addition, there are multiple sources of ethyl acetate in the
ꢀ
system: reactions of CH CH OCH(OO )CH radicals with
3
2
3
1
1 J. J. Orlando, G. S. Tyndall and T. J. Wallington, Chem. Rev.,
003, 103, 4657.
themselves or with other peroxy radicals; reaction (6) of the
-ethoxyethoxy radical with O ; and, in the presence of NO , a
2
1
2
x
12 R. E. Shetter, J. A. Davidson, C. A. Cantrell and J. G. Calvert,
Rev. Sci. Instrum., 1987, 58, 1427.
13 J. J. Orlando and G. S. Tyndall, J. Phys. Chem. A, 2002, 106, 312.
small (10–15%) occurrence of activated decomposition of
0
1
-ethoxyethoxy via H-atom elimination, reaction (4 ). Finally,
1
4 A. Notario, A. Mellouki and G. Le Bras, Int. J. Chem. Kinet.,
000, 32, 105.
a change in mechanism (lower yields of ethyl formate and ethyl
acetate) was noted at elevated temperature in the absence
2
15 M. A. Ferenac, A. J. Davis, A. S. Holloway and T. S. Dibble, J.
Phys. Chem. A, 2003, 107, 63.
of NO
x
, which might be the result of as-yet-undetermined
ꢀ
1
1
1
1
6 NIST Chemistry WebBook, http://webbook.nist.gov/chemistry/.
7 R. Atkinson, J. Phys. Chem. Ref. Data, 1994, Monograph 2, 1.
8 R. Atkinson, Int. J. Chem. Kinet., 1997, 29, 99.
9 W. Braun, J. T. Herron and D. K. Kahaner, Int. J. Chem. Kinet.,
1
chemistry of the peroxy radical, CH CH OCH(OO )CH .
3
2
3
Acknowledgements
988, 20, 51.
0 J. Arey, S. M. Aschmann, E. S. C. Kwok and R. Atkinson, J. Phys.
Chem. A, 2001, 105, 1020.
2
The National Center for Atmospheric Research is operated by
the University Corporation for Atmospheric Research, under
the sponsorship of the National Science Foundation. This
work was funded in part by the NASA ROSES (Atmospheric
Composition) program. Thanks are due to Geoff Tyndall for
many helpful discussions and comments on the manuscript,
and to Alan Fried for his careful reading of the manuscript.
21 E. M. Collins, H. W. Sidebottom, J. C. Wenger, S. Le Calve, A.
Mellouki, G. LeBras, E. Villenave and K. Wirtz, J. Photochem.
Photobiol., A, 2005, 176, 86.
2
2 T. J. Wallington, J. M. Andino, A. R. Potts, S. J. Rudy, W. O.
Siegl, Z. Zhang, M. J. Kurylo and R. E. Huie, Environ. Sci.
Technol., 1993, 27, 98.
2
3 C. Taatjes, J. Phys. Chem. A, 2006, 110, 4299.
2
4 J. Sehested, T. Mogelberg, T. J. Wallington, E. W. Kaiser and O. J.
Nielsen, J. Phys. Chem., 1996, 100, 17218.
5 M. M. Maricq, J. J. Szente and J. D. Hybl, J. Phys. Chem. A, 1997,
2
References
1
01, 5155.
1
Intergovernmental Panel on Climate Change (IPCC) (2001), Cli-
mate Change 2001: The Scientific Basis, Contribution of Working
Group I to the Third Assessment Report of the Intergovernmental
Panel on Climate Change, ed. J. T. Houghton, Y. Ding, D. J.
Griggs, M. Noguer, P. J. van der Linden, X. Da, K. Maskell and C.
A. Johnson, Cambridge University Press, New York, p. 881.
W. P. L. Carter, Documentation of the SAPRC-99 chemical
mechanism for VOC reactivity assessment, Final Report to Cali-
fornia Air Resources Board Contract No. 92-329, and (in part)
26 J. Sehested, K. Sehested, J. Platz, H. Egsgaard and O. J. Nielsen,
Int. J. Chem. Kinet., 1997, 29, 627.
27 H. J. Curran, W. J. Pitz, C. K. Westbrook, P. Dagaut, J. C.
Boettner and M. Cathonnet, Int. J. Chem. Kinet., 1998, 30, 229.
28 H. J. Curran, S. L. Fischer and F. L. Dryer, Int. J. Chem. Kinet.,
2000, 32, 741.
29 D. A. Good and J. S. Francisco, J. Phys. Chem. A, 2000, 104, 1171.
30 T. Yamada, J. W. Bozzelli and T. H. Lay, Int. J. Chem. Kinet.,
2000, 32, 435.
2
9
5-308, 2000. Available at http://pah.cert.ucr.edu/Bcarter/.
31 I. Liu, N. W. Cant, J. H. Bromly, F. J. Barnes, P. F. Nelson and B.
S. Haynes, Chemosphere, 2001, 42, 583.
3
4
5
6
7
R. G. Derwent, M. E. Jenkin, S. M. Saunders and M. J. Pilling,
Atmos. Environ., 1998, 32, 2429.
A. Mellouki, S. Teton and G. LeBras, Int. J. Chem. Kinet., 1995,
32 A. Andersen and E. A. Carter, Isr. J. Chem., 2002, 42, 245.
33 C. M. Rosado-Reyes, J. S. Francisco, J. J. Szente, M. M. Maricq
and L. F. Ostergaard, J. Phys. Chem. A, 2005, 109, 10940.
34 K. T. Kuwata, A. S. Hasson, R. V. Dickinson, E. B. Petersen and
L. C. Valin, J. Phys. Chem. A, 2005, 109, 2514.
2
7, 791.
T. J. Wallington and S. M. Japar, Environ. Sci. Technol., 1991, 25,
10.
4
J. Eberhard, C. Muller, D. W. Stocker and J. A. Kerr, Int. J. Chem.
Kinet., 1993, 25, 639.
S. A. Cheema, K. A. Holbrook, G. A. Oldershaw, D. P. Starkey
and R. W. Walker, Phys. Chem. Chem. Phys., 1999, 1, 3243.
35 F. Jorand, A. Heiss, O. Perrin, K. Sahetchian, L. Kerhoas and J.
Einhorn, Int. J. Chem. Kinet., 2003, 35, 354.
36 O. Perrin, A. Heiss, K. Sahetchian, L. Kerhoas and J. Einhorn, Int.
J. Chem. Kinet., 1998, 30, 875.
This journal is ꢄc the Owner Societies 2007
Phys. Chem. Chem. Phys., 2007, 9, 4189–4199 | 4199